Meta says it has caught up in AI and has no plans to slow down

Meta official logo (public domain, Wikimedia Commons) — CryptoBriefing brand treatment

Meta says it has caught up in AI and has no plans to slow down

Mark Zuckerberg tells the Financial Times that Meta has reached the AI frontier and will keep spending aggressively despite calls for a coordinated slowdown

Mark Zuckerberg says Meta has finished playing catch-up in artificial intelligence. In an interview with the Financial Times, the Meta chief said the company now sits at what it considers the AI frontier.

Meta does not plan to ease off the gas, even as parts of the industry push for a coordinated slowdown in AI development.

The comments come as Meta launched two new models, Muse Wonder and Muse Wonder Full, on October 12, 2026. They put Meta in direct competition with OpenAI and Anthropic, the two labs most often treated as the pace-setters in frontier AI.

The Muse lineup and the race it started

The new models extend Meta’s Muse platform. Its flagship consumer product so far has been the Muse personal agent, which launched in early 2026.

Advertisement

The Muse agent logged 6.6 million downloads and 1.8 million daily active users soon after it went live.

OpenAI responded roughly three weeks after Muse arrived, rolling out its own agent called Dots.

The leadership behind the push was expensive to assemble. Alexandr Wang serves as Meta’s chief AI officer, and Nat Friedman runs AI products. Both joined on nine-figure compensation packages.

The spending side of the equation

Meta is projected to spend between $130 billion and $145 billion in capital expenditures across 2026. Santosh Janardhan, who leads Meta’s AI infrastructure, said more than $100 billion of the budget will go toward three priorities: data centers, custom training chips, and expanding contracted nuclear power capacity to meet the company’s energy needs through 2035.

Background: from follower to frontier

Zuckerberg’s broader vision centers on what he calls personal superintelligence, meaning advanced AI that everyday consumers can use directly. Zuckerberg has published essays arguing for the distribution of both open-weight and closed-weight models.

Open-weight models release the trained parameters publicly, so anyone can download, run and modify them. Closed-weight models keep those internals private and offer access only through the company’s own products or services.

What this means

For investors, Meta’s core business is still advertising, and AI ties directly into it. Better AI can sharpen ad targeting and personalization across Meta’s platforms.

Meta’s infrastructure plans touch chip design, data center construction and energy contracts. Zuckerberg’s refusal to slow down puts Meta at odds with voices calling for a coordinated pause.

The things to watch are straightforward: whether Muse’s daily active users keep climbing, how Muse Wonder and Muse Wonder Full perform against OpenAI and Anthropic, and whether Meta’s actual spending lands within its projected range.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.
Meta says it has caught up in AI and has no plans to slow down
Meta says it has caught up in AI and has no plans to slow down

Mark Zuckerberg tells the Financial Times that Meta has reached the AI frontier and will keep spending aggressively despite calls for a coordinated slowdown

Meta official logo (public domain, Wikimedia Commons) — CryptoBriefing brand treatment

Mark Zuckerberg says Meta has finished playing catch-up in artificial intelligence. In an interview with the Financial Times, the Meta chief said the company now sits at what it considers the AI frontier.

Meta does not plan to ease off the gas, even as parts of the industry push for a coordinated slowdown in AI development.

The comments come as Meta launched two new models, Muse Wonder and Muse Wonder Full, on October 12, 2026. They put Meta in direct competition with OpenAI and Anthropic, the two labs most often treated as the pace-setters in frontier AI.

The Muse lineup and the race it started

The new models extend Meta’s Muse platform. Its flagship consumer product so far has been the Muse personal agent, which launched in early 2026.

Advertisement

The Muse agent logged 6.6 million downloads and 1.8 million daily active users soon after it went live.

OpenAI responded roughly three weeks after Muse arrived, rolling out its own agent called Dots.

The leadership behind the push was expensive to assemble. Alexandr Wang serves as Meta’s chief AI officer, and Nat Friedman runs AI products. Both joined on nine-figure compensation packages.

The spending side of the equation

Meta is projected to spend between $130 billion and $145 billion in capital expenditures across 2026. Santosh Janardhan, who leads Meta’s AI infrastructure, said more than $100 billion of the budget will go toward three priorities: data centers, custom training chips, and expanding contracted nuclear power capacity to meet the company’s energy needs through 2035.

Background: from follower to frontier

Zuckerberg’s broader vision centers on what he calls personal superintelligence, meaning advanced AI that everyday consumers can use directly. Zuckerberg has published essays arguing for the distribution of both open-weight and closed-weight models.

Open-weight models release the trained parameters publicly, so anyone can download, run and modify them. Closed-weight models keep those internals private and offer access only through the company’s own products or services.

What this means

For investors, Meta’s core business is still advertising, and AI ties directly into it. Better AI can sharpen ad targeting and personalization across Meta’s platforms.

Meta’s infrastructure plans touch chip design, data center construction and energy contracts. Zuckerberg’s refusal to slow down puts Meta at odds with voices calling for a coordinated pause.

The things to watch are straightforward: whether Muse’s daily active users keep climbing, how Muse Wonder and Muse Wonder Full perform against OpenAI and Anthropic, and whether Meta’s actual spending lands within its projected range.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.